avr-gcc-list
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [avr-gcc-list] testsuite saga continues


From: Paulo Marques
Subject: Re: [avr-gcc-list] testsuite saga continues
Date: Thu, 07 Feb 2008 20:42:38 +0000
User-agent: Thunderbird 1.5.0.14 (X11/20071210)

Wouter van Gulik wrote:
[...]
The only information you print is register info right? Since the parsing is
so heavy does it make sense to save the complete register file (up to SPH).
And then parse afterwards? It is "just" 96 bytes.
The only info missing would be addresses to/from memory, but that could be
ignored, since it's only load/store. When the address is absolutely
relevant, just re-run using log.
Just thinking out loud, it is probably nasty to create and the gain is
almost nothing...

Yes, we can play tricks to speed up the log, but the point is: we can not add a log to the non-logging version because we want that one to run as fast as possible, which by the definition of "as fast as possible" excludes logging data ;)

Having a logging version that outputs only the last N lines of the log at exit time, will already make the logging version faster. So, I would start there and grab the low hanging fruit first.

[...]
I don't have those numbers right now, but since there are tests that
don't even fit in 128Kb of flash, there are probably some more that
don't fit on 8Kb.

Aha, well it is going to be hard to test the <8Kb then. Are these 128KB even
in an optimized? I can imagine not fitting when using -O0.

The ones that come to my mind are the compat/struct-by-value*. These are horrible constructs of functions that receive structures by value. They seem to be compiled without optimizations options which is equivalent to -O0.

At one point I was able to see the size of one of these monsters, and it had a .text section larger than 400Kb. So, I don't think these really matter much for avr. Even the testsuite marks these as unsupported and not as failed.

So we probably can run the testsuite with the smaller architectures, and the only downside will be a few more unsupported / unresolved test cases. I might give it a try later,

--
Paulo Marques
Software Development Department - Grupo PIE, S.A.
Phone: +351 252 290600, Fax: +351 252 290601
Web: www.grupopie.com

"The impossible often has a kind of integrity to it which the merely
 improbable lacks."
Douglas Adams




reply via email to

[Prev in Thread] Current Thread [Next in Thread]